Analysis of Camera Pose Estimation Using 2D Scene Features for Augmented Reality Applications

Augmented reality (AR) enables civil engineers and architects to visualize a representation of the structure that is going to be built at different stages of the construction. Overlaying a 3D model onto an image requires localizing the camera accurately in its environment. In this paper, we evaluate the camera pose estimation methods using circles and straight lines, as two of the common features visible in the architectural structures, by taking into account the relationship between the coordinates of the features in the 2D image and their corresponding positions in the 3D world. The proposed approach could be used for AR applications.